Search⌘ K
Home
Courses
Concurrent Data Processing in Elixir

Register New Processes and Query the Registry

Learn to register new processes and query the registry using the select function.

We'll cover the following...

Register new processes

We make a few changes to job.ex. First, let’s add a helper function, via/2, to job.ex:

C++
#file path -> jobber/lib/jobber/job.ex
defp via(key, value) do
  {:via, Registry, {Jobber.JobRegistry, key, value}}
end

This returns a tuple that conforms to the required naming specification. The format is always {:via, Registry, config}, where config is also a tuple and can either be {registry_process, key} or {registry_process, key, value}. We use the extra value element to store some useful metadata about the process. We label each job with a type description so it’s easier to find out what each process is doing.

Modify start_link

We can use the provided ID or generate one for the process’s name. Let’s move some of the existing logic in start_link/1:

C++
#file path -> jobber/lib/jobber/job.ex
def start_link(args) do
  args =
    if Keyword.has_key?(args, :id) do
      args
    else
      Keyword.put(args, :id, random_job_id())
    end
  id = Keyword.get(args, :id)
  type = Keyword.get(args, :type)
  GenServer.start_link(__MODULE__, args, name: via(id, type))
end

We add an extra :name argument to GenServer.start_link/2 to name the process. We have to provide the :via tuple, generated by the via/2 helper function. We ...